Thomas Nierobisch

Large View Visual Servoing of a Mobile Robot with a Pan-Tilt Camera

Vision being the most important human sensor and feedback system is considered to play a prominent role in the future of robotics. Robust techniques for visual robot navigation are essential in service robotics tasks. This contribution presents a novel approach to large view visual servoing of a mobile robot with a pan-tilt camera. A predictive camera control scheme guarantees that the image features remain in the cameras field of view. In a demonstration the robot is manually guided through the environment and reference images of landmarks are taken at distinct waypoints. During autonomous navigation the robot later follows the demonstrated path solely based on the visual feedback provided by the landmarks. The camera gaze is controlled independent of the robots egomotion which enables the robot to track landmark features across larger distances and orientations. Therefore, fewer visual landmarks are required to describe a smooth path through the environment. In order to control the robot motion independent of the gaze the observed features are transformed from the image plane into a virtual camera plane. The rotational, lateral and longitudinal motion are controlled separately by selecting appropriate image features that decouple the rotational and translational velocity component. This property is particular useful for traversing confined indoor spaces such as corridors, in which landmarks are located lateral to the direction of robot motion

Vibration control of a multi-link flexible robot arm with Fiber-Bragg-Grating sensors

Flexible, lightweight manipulators offer some advantages in contrast to rigid arms, such as compact and lighter drives, energy efficiency, reduced masses and costs. This paper presents a novel approach for vibration damping of a multi-link flexible arm. The strain of the elastic arms is measured with Fiber-Bragg-Grating (FBG) sensors and provides the feedback signal to dampen their flexural dynamics. A dynamic model of a three link arm is derived that accounts for the rigid and flexural dynamics including gravity. The arm vibrations are damped by nonlinear strain feedback. The controller is general and robust and its design does not require a model of the flexural dynamics. In the context of closed loop vibration control FBG sensors offer a better signal to noise ratio compared to strain gauges, which allows a higher static gain in the feedback loop with more efficient dissipation of vibrational energy. The feasibility and effectiveness of the proposed vibration control scheme in conjunction with FBG sensors is verified and analyzed in simulations and confirmed in experiments with a flexible three link robot arm.

Optimal Large View Visual Servoing with Sets of SIFT Features

This paper presents a novel approach to large view visual servoing in the context of object manipulation. In many scenarios the features extracted in the reference pose are only perceivable across a limited region of the work space. The limited visibility of features necessitates the introduction of additional intermediate reference views of the object and requires path planning in view space. In our scheme the visual control is based on decoupled moments of SIFT-features, which are generic in the sense that the control operates with a dynamic set of feature correspondences rather than a static set of individual features. The additional freedom of dynamic feature sets enables flexible path planning in the image space and online selection of optimal reference views during servoing to the goal view. The time to convergence to the goal view is estimated by a neural network based on the residual feature error and the quality of the SIFT feature distribution. The transition among reference views occurs on the basis of this estimated cost which is evaluated online based on the current set of visible features. The dynamic switching scheme achieves robust and nearly time-optimal convergence of the visual control across the entire task space. The effectiveness and robustness of the scheme is confirmed in an evaluation in a virtual reality simulation and on a real robot arm with a eye-in-hand configuration.

Two-Finger Grasping for Vision Assisted Object Manipulation

The problem of object manipulation with no prior knowledge of object pose and geometry constitutes a key problem in the domain of service robotics. The grasping methodologies reported so far use either force closure or friction between the fingers and the object surfaces. Force closure leads to stable grasping such that the total sum of forces acting upon the object becomes zero [1]. Grasping methods based upon friction are better suited for grippers with limited degrees of freedom, for which force closure is infeasible due to gripper geometry. The fundamental objective is to maximize the contact area and thereby the friction force between the gripper and the object by selecting suitable contact points. In order to determine optimal contact points, [2] and [3] proposed vision based grasp point selection from the extracted contours of the object. Subsequently, the force applied to the object is determined based upon the selected grasp point and object properties. The force required for stable grasping is difficult to predict due to the presence of friction in gripper actuators as well as the nonlinear behavior of gripper and object materials. Conventional control strategies are not suitable for such conditions. Therefore [4] investigates a fuzzy logic approach for force control. Reference [5] addresses this issue from a different perspective as slippage is detected by means of dynamic tactile sensing.

Bildbasierte Navigation eines mobilen Roboters mittels omnidirektionaler und schwenkbarer Kamera

Dieser Beitrag prasentiert einen neuartigen Ansatz zur entkoppelten Regelung der Kamera-Blickrichtung und der Bewegung eines mobilen Roboters im Kontext der bildbasierten Navigation. Eine schwenkbare monokulare Kamera halt unabhangig von der Roboterbewegung die relevanten Merkmale fur die Navigation im Sichtfeld. Die Entkopplung der Kamerablickrichtung von der eigentlichen Roboterbewegung wird durch die Projektion der Merkmale auf eine virtuelle Bildebene realisiert. In der virtuellen Bildebene hangt die Auspragung der visuellen Merkmale fur die bildbasierte Regelung nur von der Roboterposition ab und ist unabhangig gegenuber der tatsachlichen Blickrichtung der Kamera. Durch die Schwenkbarkeit der monokularen Kamera wird der Arbeitsbereich, uber dem sich ein Referenzbild zur bildbasierten Regelung eignet, gegenuber einer statischen Kamera signifikant vergrosert. Dies ermoglicht die Navigation auch in texturarmen Umgebungen, die wenig verwertbare Textur- und Strukturmerkmale aufweisen.

Time-optimal large view visual servoing with dynamic sets of SIFT

This paper presents a novel approach to large view visual servoing in the context of object manipulation. In many scenarios the features extracted in the reference pose are only perceivable across a limited region of the work space. The limited visibility of features necessitates the introduction of additional intermediate reference views of the object and requires path planning in view space. In our scheme visual control is based on decoupled moments of SIFT-features, which are generic in the sense that the control operates with a dynamic set of feature correspondences rather than a static set of geometric features. The additional flexibility of dynamic feature sets enables flexible path planning in the image space and online selection of optimal reference views during servoing to the goal view. The time to convergence to the goal view is estimated by a neural network based on the residual feature error and the quality of the SIFT feature distribution. The transition among reference views occurs on the basis of this estimated cost which is evaluated online based on the current set of visible features. The dynamic switching scheme achieves robust and nearly time-optimal convergence of the visual control across the entire task space. The effectiveness and robustness of the scheme is confirmed in an experimental evaluation in a virtual reality simulation and on a real robot arm with a eye-in-hand configuration.